Abstract— Autism Spectrum Disorder

Abstract— Autism Spectrum Disorder (ASD) is one of the major developmental disorders in terms of neurodevelopment brain functions. The three symptoms that an autistic patient shows are: social development, communication problem and odd social behavior. Proposed paper presents the design and development of an application intended for the detection and diagnosis of autism in children at the early stage. The application is of tremendous help in detection, diagnosis and monitoring the health and treatment of patient. This work presents an architectural solution that enables the operability between an autistic child and the doctor to contribute in the treatment. The solution is implemented by developing an android mobile application.
The survey results showed that the application made a positive treatment contribution to a person with ASD. Also the monitoring system was improvised with tracking.

Keywords— autism; learning; schedules; games; monitoring; mobile application

I. INTRODUCTION

One of the major mental disorder that exists today is Autism which impairs capturing and gaining capabilities and it difficult to interact socially. In India, 8.1% children of age group between 2-12 years are considered to have ASD syndromes. This neurological mental disorder gives symptoms of unusual restricted and repetitive behaviors. It is a neurodevelopmental disorder of brain function and not a disease. It is often associated with brain disability and difficulties in mind-body coordination and attention deficiency and sometimes in physical health issues that involve sleep and digestion. Early signs of autism develop from the tender age of 2 to 3 years. This is a crucial age where preventable measures can be taken to avoid further complications in the condition. However, later signs of autism may be detected through behavioral reactions and reading apt or social interactions while meeting strangers. In few cases, autistic children may reach their development stage normally after showing few signs of improvement, however there are still chances of relapse .One of the biggest drawbacks for an Autistic person is the inability to converse and communicate their needs in general with the people. In some cases, autistic children doesn’t even manage to develop their natural speech, due to which there is very little scope for improvement in their communication skills. Therefore diagnosis at a younger age is beneficial for treatment.
Communication skills lacks in Autistic children since they feel more comfortable interacting and communicating with their loved ones as compared to acquaintances. This leads to a problem in terms of developing their learning skills especially when learning requires interaction and familiarity between the teacher and the autistic child. Due to the necessity of one-on-one interaction, education poses as a major challenge for an autistic person where particular techniques and tasks are required to be done towards building conceptual and operational skills. This work identifies particular features that use the current enabled technologies existing now-a-days (such as mobile application devices and cloud computing) that can contribute in the detection, treatment (in the form of learning) and monitoring (both treatment assessment, progress and the physical and mental condition) of the autistic child. In this paper we propose an architectural design of the overall system and presents the complete implementation of the mobile application.
The terms Autism and ASD (Autism Spectrum Disorders) can be interchangeable in the medical setting, and have been used to describe one of the most intriguing neurobehavioral syndromes, that include the so-called “triad of Wing”: problems in communication, social skills, and restrict repertoire of interests. However, it is somewhat difficult to precisely define autism, because of the imprecise boundaries between different kinds of ASD.
The general management of ASD from the clinical perspective encompasses both interventions in the family/environment as well as interventions addressed to the patient. Ideally, after Recent Advances in Autism Spectrum Disorders – Volume I 634 diagnosis confirmation, the best initial approach could be done by an interdisciplinary team including professionals coming from medicine, psychology and social sciences. Obviously, before initiating any kind of intervention, several steps must be done as follows. First of all, the final diagnosis must be confirmed by a careful anamnesis as well double-checked using the DSM-IV criteria as well as a reliable clinical instrument such as Autism Diagnosis Interview-Revised (ADI-R) (Becker et al. 2012). The ADI-R is frequently used as a gold standard instrument for publication purposes, but it is problematic in the clinical practice for several reasons, such as it can miss same ASD cases as well as it need at least two hours to be completed. Then, the intensity of the ASD could be defined both from the clinical perspective and by one instrument such as CARS (Pereira, Riesgo, and Wagner 2008). Another critical issue is to delimitate if there is any associated mental disability and its degree of intensity. As clinicians, we know the prognostic importance of an unaffected intelligence in ASD patients.
The second step includes the definition of the parent’s doubts, fears, and degree of awareness.
Generally, after diagnosis confirmation, parents become stressed. Not infrequently they go to internet in order to search every kind of available information regarding autism. Because some information coming from internet can be inaccurate, at this point, it is very important to clarify which are the evidence-based types of therapies to date.
The third step could be the delimitation of environmental variables that needs to be addressed, starting from the home and family. Neighborhood and school needs to be evaluated both in terms of potential stressors and also because they can facilitates choosing a given type of therapy on an individual basis.
The next step is done by the identification of the target behaviors needing treatment. After core symptoms definition in each case, the different professional specialties that need to be involved are selected. In general, the team includes a doctor specialized in ASD patients as well as one speech therapist and others professionals arising from health care and/or education with experience in children with ASD.

II. LITERATURE SURVEY

Below is the literature survey on the papers about detection, diagnosis and clinical treatment alternatives for autism in children.

1 Machine learning for early detection of autism using a parental questionnaire

This paper concentrates on the detection of autism in children in the first few years of their life using machine learning algorithm. They used Machine Learning to gold standard clinical data obtained from thousands of children at risk for autism spectrum disorders to build an economic, fast, and easy to apply autism screening tool that performs better than most widely used standardized instruments.
The autism machine learning based testers were trained using information compiled from different clinical repositories of ADOS and ADI-R score-sheets of nearly 7 thousand children between 18 to 84 months of birth, supplemented by data collected from the guardians of children answering screening questions on Cognoa’s website. Both testers were applied using Cognoa’s Application in a medical trial to a sample of 162 at-risk children.
The parental questionnaire tester aims on behavioral patterns typically present in a standard autism diagnostic instrument, the Autism Diagnostic Interview Revised (ADI-
R) .This clinical tool consists of a guardian interview of
93 multi-part questions with multiple choice which are delivered by a trained professional in a clinical environment.

2 Deep Learning Based Recognition of Meltdown in Autistic Kids

This paper basically concentrates on the recognition of meltdown in autistic children. Kids with autism generally experience sudden meltdowns which not only makes the situation difficult for the guardian but also make the kids hurt themselves physically. Researchers have found that children with autistic spectrum disorder exhibit certain actions using which we can predict meltdowns in the children.
They first went through the process of collecting raw pictures and videos that contained the desired actions. This data was attained from multiple databases and reliable internet sources, which is followed by the data cleaning process that crucially involved filtering the database from deplorable pictures and frames thereby attaining a more feasible set of images for training which is further processed by cropping and annotating using multiple picture processing tools.
Around 2000 data images were assembled for each class of behavior, and 92% classification accuracy was attained for training alone .Training time taken in class was approximately around 10-15 minutes.

3 Analysis of Quantitative Autistic Traits While Teaching Social skills to ASD Children Using Multimedia Tool

This paper gives quantitative analysis of autistic traits and evaluates technology based training for upcoming cognitive skills in ASD children. It mainly explores the social competence in learning environment and attempts to strengthen behavioral performances in cognitive and social interactions. It was observed that ASD children after random trials became less impulsive, restless and more trivial.
The increase in mood constancy was also observed .Multimedia intrusion is found to be greatly successful in treating problems like autism. In the survey it came to appearance that In India, 8.1% children of age group of 4-14 years were predicted to have ASD. This neurological disorder gives symptoms of unusual restricted and repetitive behaviors. Other symptoms observed were concentration issues and maintain eye contact and focus in particular activity in addition to reading, writing and emotional imbalance.
In India the early signs of ASD are usually ignored and hard to accept by the family due to various reasons like social acceptance of the problem, lack of knowledge and short of finances. The delay in detection and diagnosis of the problem leads to overlapping of multiple medical conditions along with ASD which poses a problem for special educators and parents to finally get a way to teach ASD children and control their behaviors. Due to fast pace changing lifestyle time has become a restriction and limiting social interactions. There are so many family issues in considerations that can be the reason why the parents fail to give productive time to teach different cultures and customs that exist in society to their children. Apart from the family issues of autistic children to be taken into consideration, few issues are still a point of concern like providing ASD children with regular activities to engage and control their temperament.
There are different teaching mechanisms available to teach autistic children, very effective and prominent method of teaching is RPM by Soma 1 .{ 1 Rapid Prompting methodology http://www.halo-soma.org}
Multimedia is the combination of collection and easily available contents such as message, audio, images, text, animation, video and interactive materials. Multimedia differs from media that use only basic computer display such as text or conventional forms of symbols and clip arts or hand-made material.
Multimedia can be used as per the requirement likewise it can be recorded ,displayed and retrieved with the help of computerized and electronic devices, data can be captured on real time devices as well and do live recording also. It was observed that autistic children have better response to pictures and image recognition. One hour interactive session was conducted with the help of therapists and special educator, evaluations were directed depending on the child’s nature, age group and interest.
As an interruption from further activities, the child was given a chance to interact with a mobile, desktop and with h sequences of commercially available apps, movies and short stories for preschooler’s, and music to calm their anxiety level. It was observed that all the children were able to interact with the gadgets and trying to learn and explore the touch screen interface. After observing children of designated parameters for assessment measures, stratified random allocation was completed and parents/therapists informed by reports generated via weekly meets and through telephonic conversation. The results obtained after intervention are indicated in the figure. Here ZC indicates No Change, NC indicates Negative Change and PC indicate Positive change observed after intervention using multimedia tools.

4 AutiAid: A Learning Mobile Application for Autistic Children
This paper presents an architectural solution that enables the operability between an autistic child under parental monitor system and doctor associated to contribute on the early diagnosis, detection, treatment and monitoring to persons with ASD. The solution is implemented by creating a mobile android application and tested with a population with Autism syndromes.
The results show that the application made a positive diagnosis and treatment contribution to a person with ASD. Obvious signs of autism develop from the tender age of 2 to 3 years. However, later signs of autism may be detected through repetitive behavioral actions and readings or social interactions while meeting strangers. In some cases, autistic children may reach their development age normally after exhibiting signs of improvement; however, there are still chances of relapse in the progress.
Communication skills lacks in Autistic persons since they feel more comfortable in interacting and communicating with their known ones as compared to strangers. This leads to a problem in terms of developing their learning skill especially when learning requires communication and familiarity between the teacher and the autistic child (person). This work identifies specific features that uses the current existential technology that exist which can contribute in the treatment (in the form of learning) and monitoring (both treatment progress and the physical location) of the autistic child. In this paper, an architectural and component design of the overall system and present the complete implementation.
DISSERO application was introduced to improve mental and social skills, shows how the learning process may be developed along with apt media attention towards needs. . This paper contains following sections presenting related work, the functional requirements, the system design and implementation, the experimental results captured from this work, and conclusions were drawn from various inferences were made.

III PROPOSED SYSTEM

The architectural design proposes a systematic organization to detect the early signs of ASD in children. Clinically to detect the presence of autism disorder in children two broad spectrums are defined and suitable scale is decided accordingly. CARS (Childhood autism rating scale) is conducted for children of age above 3 years and M-CHAT (Modified checklist for autism tolerance) is conducted for children of age below 3 years. Level of ASD is recognized by DSM-V where the DSM-V is the standard reference that healthcare providers use to diagnose mental and behavioral conditions, including autism. Levels of ASD recognized by DSMV are, Level 1: Requiring support (Mild) Poor social skills, difficulty in initiating interactions, attempts to make friends are odd and unsuccessful. Level 2: Requiring substantial support (Moderate) Markedly odd, restricted repetitive behaviors, noticeable difficulties changing activities or focus. Level 3: Requiring Very Substantial Support (Severe) Very limited speech, odd, repetitive behavior; many express their basic needs only. It is also important to keep in consideration that the IQ level of the patient can vary depending on the level of severity.
Based on the aforementioned symptoms the proposed functional system is divided into four main subsystems –Account system, Schedule system, Monitoring system and Progress section. The Account section enables the users to register as-well as sign-in. The users recognized here are the caretaker (or guardian), associated doctors and the patient (child). The account system also manages accounts registered within it and deals with associating supervisors, and caregiver to a specific autistic child.

SCHEDULE AND MONITORING SYSTEM

The schedule system is composed of three types of schedules: Medical, Learning and Daily schedules. The medical section consists of the patient’s past medical record inferences and treatment undertaken and further diagnosis and treatment tracks of the future. The Medical Schedule also relates to allocating the doctor to see an autistic child at a specific time. The learning schedule allocates a time for the autistic child to take learning assignments under parental guidance. The schedule system also enables the doctor to monitor the progress of the child and accordingly modify the curriculum/course of activities performed by the autistic child. The doctor can also design his own course of activities and create a schedule for the patient accordingly.
The autistic child can remember images better than words thus using it towards building a daily routine for the child can prove beneficial in the learning process. Manipulating these skill, leads children to better understanding and memorizing of the words. For this use, it is proposed that an application be built where a list of words are provided along with corresponding images for relational understanding between the images and their meaning. It is crucial to provide patients with new and updated images to broaden their learning spectrum therefore it is essential to keep a rotation of the displayed images. The doctor can view the results of the patients’ exercises and create and assign new exercises to him. Both the parents and the doctor can communicate with one another using the application and then decide the best course of actions for the autistic patient.